Research map / Label-efficient and robust learning / Test-time adaptation
Learning Topology-Aware Representations via Test-Time Adaptation for Anomaly Segmentation
Ali Zia, Usman Ali, Abdul Rehman, Umer Ramzan, Kang Han, Muhammad Faheem, Shahnawaz Qureshi, Wei Xiang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use TopoTTA instead of THR thresholding | PatchCore · MVTec AD | F1: 0.136 -> 0.553 | Improved |
| use TopoTTA instead of TTT4AS | PatchCore · MVTec AD | F1: 0.382 -> 0.553 | Improved |
| use TopoTTA instead of TTT4AS | Dinomaly · VisA | F1: 0.267 -> 0.470 | Improved |
| use TopoTTA instead of TTT4AS | M3DM · MVTec 3D-AD | F1: 0.468 -> 0.482 | Improved |
| use TopoTTA instead of TTT4AS | PO3AD · AnomalyShapeNet | F1: 0.493 -> 0.504 | Improved |
| use TopoTTA instead of THR | MambaAD · Real-IAD | F1: 0.228 -> 0.319 | Improved |
| use only the Top1 persistence component rather than Top5 (ablation) | TopoTTA with PatchCore · MVTec AD | F1: 0.370 -> 0.553 | Improved |
| use Top1 rather than Top2 persistence component (ablation) | TopoTTA with M3DM · MVTec 3D-AD | F1: 0.336 -> 0.482 | Improved |
| increase K from 1 to 5 (ablation) | TopoTTA with PatchCore · MVTec AD | recall: 0.851 -> 0.912 | Improved |
| add superlevel filtration to sublevel with EAI (ablation) | TopoTTA with PatchCore · MVTec AD | F1: 0.370 -> 0.553 | Improved |
| use superlevel only instead of combined filtration (ablation) | TopoTTA with CMM · MVTec 3D-AD | precision: 0.447 -> 0.082 | Hurt |
| add PCES contrastive encoder (ablation) | TopoTTA with PatchCore · MVTec AD | F1: 0.423 -> 0.553 | Improved |
Closest papers by meaning
- Topology-Aware Optimal Transport for Source-Free Test-Time Adaptation in Anomaly Segmentation (NeurIPS 2026 · similarity 0.95)
- TopoTTA: Topology-Enhanced Test-Time Adaptation for Tubular Structure Segmentation (ICCV 2025 · similarity 0.88)
- Topograph: An Efficient Graph-Based Framework for Strictly Topology Preserving Image Segmentation (ICLR 2025 · similarity 0.85)
- Towards Scalable Topological Regularizers (ICLR 2025 · similarity 0.85)
- TTVD: Towards a Geometric Framework for Test-Time Adaptation Based on Voronoi Diagram (ICLR 2025 · similarity 0.84)
- FAST: Foreground‑aware Diffusion with Accelerated Sampling Trajectory for Segmentation‑oriented Anomaly Synthesis (NeurIPS 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
